TCP输入处理是系统中最长的一部分代码,tcp_input函数约有1100行代码(预警!)。它完全遵循RFC793中定义的输入事件处理步骤,这些步骤详细定义了如何根据连接的当前状态,处理不同的输入报文段。
当发现分组IP首部中的协议字段是TCP协议时,IP协议的软中断处理函数ipintr调用tcp_input函数进行处理。tcp_input函数我删去了处理URG标志的流程,它的代码如下:
/*
* TCP input routine, follows pages 65-76 of the
* protocol specification dated September, 1981 very closely.
*/
void
tcp_input(m, iphlen)
register struct mbuf *m;
int iphlen;
{
register struct tcpiphdr *ti;
register struct inpcb *inp;
u_char *optp = NULL;
int optlen;
int len, tlen, off;
register struct tcpcb *tp = 0;
register int tiflags;
struct socket *so;
int todrop, acked, ourfinisacked, needoutput = 0;
short ostate;
struct in_addr laddr;
int dropsocket = 0;
int iss = 0;
u_long tiwin, ts_val, ts_ecr;
int ts_present = 0;
tcpstat.tcps_rcvtotal++;
/*
* Get IP and TCP header together in first mbuf.
* Note: IP leaves IP header in first mbuf.
*/
ti = mtod(m, struct tcpiphdr *);
if (iphlen > sizeof (struct ip)) /*IP首部包含选项?*/
ip_stripoptions(m, (struct mbuf *)0); /*过滤掉IP首部选项*/
if (m->m_len < sizeof (struct tcpiphdr)) { /*IP首部和TCP首部不在一个mbuf?*/
if ((m = m_pullup(m, sizeof (struct tcpiphdr))) == 0) { /*将IP首部和TCP首部调整到一个mbuf中*/
tcpstat.tcps_rcvshort++;
return;
}
ti = mtod(m, struct tcpiphdr *);
}
/*
* Checksum extended TCP header and data.
*/
tlen = ((struct ip *)ti)->ip_len; /*经过IP层处理后,tlen是TCP报文段的长度*/
len = sizeof (struct ip) + tlen; /*IP数据报总长度*/
ti->ti_next = ti->ti_prev = 0;
ti->ti_x1 = 0;
ti->ti_len = (u_short)tlen;
HTONS(ti->ti_len);
if (ti->ti_sum = in_cksum(m, len)) { /*验证校验和失败,丢弃报文*/
tcpstat.tcps_rcvbadsum++;
goto drop;
}
/*
* Check that TCP offset makes sense,
* pull out TCP options and adjust length. XXX
*/
off = ti->ti_off << 2; /*TCP首部的长度*/
if (off < sizeof (struct tcphdr) || off > tlen) { /*无效的TCP首部,丢弃*/
tcpstat.tcps_rcvbadoff++;
goto drop;
}
tlen -= off; /*数据的长度*/
ti->ti_len = tlen;
if (off > sizeof (struct tcphdr)) { /*TCP首部包含选项?*/
if (m->m_len < sizeof(struct ip) + off) { /*TCP首部和选项不在一个mbuf?*/
if ((m = m_pullup(m, sizeof (struct ip) + off)) == 0) { /*将IP首部和TCP首部调整到一个mbuf中*/*/
tcpstat.tcps_rcvshort++;
return;
}
ti = mtod(m, struct tcpiphdr *);
}
optlen = off - sizeof (struct tcphdr); /*TCP选项的长度*/
optp = mtod(m, u_char *) + sizeof (struct tcpiphdr); /*指向选项*/
/*
* Do quick retrieval of timestamp options ("options
* prediction?"). If timestamp is the only option and it's
* formatted as recommended in RFC 1323 appendix A, we
* quickly get the values now and not bother calling
* tcp_dooptions(), etc.
*/
if ((optlen == TCPOLEN_TSTAMP_APPA ||
(optlen > TCPOLEN_TSTAMP_APPA &&
optp[TCPOLEN_TSTAMP_APPA] == TCPOPT_EOL)) &&
*(u_long *)optp == htonl(TCPOPT_TSTAMP_HDR) &&
(ti->ti_flags & TH_SYN) == 0) { /*选项中只包含时间戳选项?*/
ts_present = 1;
ts_val = ntohl(*(u_long *)(optp + 4)); /*记录对端的时间戳*/
ts_ecr = ntohl(*(u_long *)(optp + 8)); /*记录本端回显的时间戳*/
optp = NULL; /* we've parsed the options */
}
}
tiflags = ti->ti_flags; /*获取TCP首部中的标志*/
/*
* Convert TCP protocol specific fields to host format.
*/
NTOHL(ti->ti_seq);
NTOHL(ti->ti_ack);
NTOHS(ti->ti_win);
NTOHS(ti->ti_urp);
/*
* Locate pcb for segment.
*/
findpcb:
inp = tcp_last_inpcb;
if (inp->inp_lport != ti->ti_dport ||
inp->inp_fport != ti->ti_sport ||
inp->inp_faddr.s_addr != ti->ti_src.s_addr ||
inp->inp_laddr.s_addr != ti->ti_dst.s_addr) {
inp = in_pcblookup(&tcb, ti->ti_src, ti->ti_sport,
ti->ti_dst, ti->ti_dport, INPLOOKUP_WILDCARD); /*根据源IP地址、源端口号、目的IP地址和目的端口号
查找合适的inpcb结构*/
if (inp)
tcp_last_inpcb = inp; /*记录最新查找到的inpcb*/
++tcpstat.tcps_pcbcachemiss;
}
/*
* If the state is CLOSED (i.e., TCB does not exist) then
* all data in the incoming segment is discarded.
* If the TCB exists but is in CLOSED state, it is embryonic,
* but should either do a listen or a connect soon.
*/
if (inp == 0) /*找不到inpcb,丢弃报文,并发送RST*/
goto dropwithreset;
tp = intotcpcb(inp);
if (tp == 0) /*没有tcpcb?丢弃报文,并发送RST*/
goto dropwithreset;
if (tp->t_state == TCPS_CLOSED) /*连接状态为CLOSED,丢弃报文*/
goto drop;
/* Unscale the window into a 32-bit value. */
if ((tiflags & TH_SYN) == 0) /*计算通告的窗口大小*/
tiwin = ti->ti_win << tp->snd_scale;
else
tiwin = ti->ti_win;
so = inp->inp_socket;
if (so->so_options & SO_ACCEPTCONN) { /*该socket是监听socket?
SO_ACCEPTCONN标志在执行listen系统调用时被置上!*/
if ((tiflags & (TH_RST|TH_ACK|TH_SYN)) != TH_SYN) { /*收到的不是纯SYN报文?*/
/*
* Note: dropwithreset makes sure we don't
* send a reset in response to a RST.
*/
if (tiflags & TH_ACK) { /*带ACK标志的无效的SYN报文,丢弃,并发送RST*/
tcpstat.tcps_badsyn++;
goto dropwithreset;
}
goto drop; /*丢弃报文*/
}
so = sonewconn(so, 0); /*创建一个新的socket!*/
if (so == 0)
goto drop;
/*
* This is ugly, but ....
*
* Mark socket as temporary until we're
* committed to keeping it. The code at
* ``drop'' and ``dropwithreset'' check the
* flag dropsocket to see if the temporary
* socket created here should be discarded.
* We mark the socket as discardable until
* we're committed to it below in TCPS_LISTEN.
*/
dropsocket++;
inp = (struct inpcb *)so->so_pcb;
inp->inp_laddr = ti->ti_dst; /*设置新socket的本地地址*/
inp->inp_lport = ti->ti_dport; /*设置新socket的本地端口*/
tp = intotcpcb(inp);
tp->t_state = TCPS_LISTEN; /*连接状态设置为TCPS_LISTEN!会执行329行的流程*/
/* Compute proper scaling value from buffer space
*/
while (tp->request_r_scale < TCP_MAX_WINSHIFT && /*根据接收缓冲区大小计算窗口缩放因子*/
TCP_MAXWIN << tp->request_r_scale < so->so_rcv.sb_hiwat)
tp->request_r_scale++;
}
/*
* Segment received on connection.
* Reset idle time and keep-alive timer.
*/
tp->t_idle = 0; /*连接空闲时间清零*/
tp->t_timer[TCPT_KEEP] = tcp_keepidle; /*保活定时器复位*/
/*
* Process options if not in LISTEN state,
* else do it below (after getting remote address).
*/
if (optp && tp->t_state != TCPS_LISTEN) /*解析不是LISTEN状态的TCP选项*/
tcp_dooptions(tp, optp, optlen, ti,
&ts_present, &ts_val, &ts_ecr);
/*
* Header prediction: check for the two common cases
* of a uni-directional data xfer. If the packet has
* no control flags, is in-sequence, the window didn't
* change and we're not retransmitting, it's a
* candidate. If the length is zero and the ack moved
* forward, we're the sender side of the xfer. Just
* free the data acked & wake any higher level process
* that was blocked waiting for space. If the length
* is non-zero and the ack didn't move, we're the
* receiver side. If we're getting packets in-order
* (the reassembly queue is empty), add the data to
* the socket buffer and note that we need a delayed ack.
*/
/*首部预测算法,tcp_input函数的一条快速路径!*/
if (tp->t_state == TCPS_ESTABLISHED && /*连接处于ESTABLISHED状态*/
(tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) == TH_ACK && /*只有ACK标志*/
(!ts_present || TSTMP_GEQ(ts_val, tp->ts_recent)) && /*没有时间戳或是新的时间戳*/
ti->ti_seq == tp->rcv_nxt && /*序列号是期望收到下一个数据的序列号*/
tiwin && tiwin == tp->snd_wnd && /*对端通告的窗口大小没有变化*/
tp->snd_nxt == tp->snd_max) { /*上一个报文段不是重传报文*/
if (ts_present && TSTMP_GEQ(ts_val, tp->ts_recent) &&
SEQ_LEQ(ti->ti_seq, tp->last_ack_sent)) { /*哪个时间戳需要回显,正确的算法!!!!*/
tp->ts_recent_age = tcp_now; /*最近一次ts_recent被更新的时间戳*/
tp->ts_recent = ts_val; /*对端发送的最新的有效时间戳*/
}
if (ti->ti_len == 0) { /*不包含任何数据*/
if (SEQ_GT(ti->ti_ack, tp->snd_una) &&
SEQ_LEQ(ti->ti_ack, tp->snd_max) &&
tp->snd_cwnd >= tp->snd_wnd) { /*有效的纯ACK报文段*/
/*
* this is a pure ack for outstanding data.
*/
++tcpstat.tcps_predack;
if (ts_present) /*如果有时间戳,根据时间戳就可计算报文的RTT,然后更新RTO*/
tcp_xmit_timer(tp, tcp_now-ts_ecr+1);
else if (tp->t_rtt && /*如果没有时间戳,根据t_rtt计算报文的RTT,然后更新RTO*/
SEQ_GT(ti->ti_ack, tp->t_rtseq))
tcp_xmit_timer(tp, tp->t_rtt);
acked = ti->ti_ack - tp->snd_una; /*ACK已确认的字节数*/
tcpstat.tcps_rcvackpack++; /*更新统计*/
tcpstat.tcps_rcvackbyte += acked;
sbdrop(&so->so_snd, acked); /*丢弃发送缓冲区中已被确认的数据*/
tp->snd_una = ti->ti_ack; /*更新snd_una指针*/
m_freem(m); /*没有数据,可释放该mbuf*/
/*
* If all outstanding data are acked, stop
* retransmit timer, otherwise restart timer
* using current (possibly backed-off) value.
* If process is waiting for space,
* wakeup/selwakeup/signal. If data
* are ready to send, let tcp_output
* decide between more output or persist.
*/
if (tp->snd_una == tp->snd_max) /*没有数据需要被确认?重传定时器复位*/
tp->t_timer[TCPT_REXMT] = 0;
else if (tp->t_timer[TCPT_PERSIST] == 0) /*否则不是持续状态?启动重传定时器*/
tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
if (so->so_snd.sb_flags & SB_NOTIFY) /*唤醒在发送缓冲区等待的进程*/
sowwakeup(so);
if (so->so_snd.sb_cc) /*还有待发送的数据,调用tcp_output发送*/
(void) tcp_output(tp);
return;
}
} else if (ti->ti_ack == tp->snd_una && /*正常数据报文段,ACK标志置位,但是未确认任何数据*/
tp->seg_next == (struct tcpiphdr *)tp &&
ti->ti_len <= sbspace(&so->so_rcv)) {
/*
* this is a pure, in-sequence data packet
* with nothing on the reassembly queue and
* we have enough buffer space to take it.
*/
++tcpstat.tcps_preddat;
tp->rcv_nxt += ti->ti_len; /*更新rcv_nxt*/
tcpstat.tcps_rcvpack++;
tcpstat.tcps_rcvbyte += ti->ti_len;
/*
* Drop TCP, IP headers and TCP options then add data
* to socket buffer.
*/
m->m_data += sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr); /*丢弃IP首部,TCP首部*/
m->m_len -= sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
sbappend(&so->so_rcv, m); /*将数据添加到接收缓冲区*/
sorwakeup(so); /*唤醒在接收缓冲区等待的进程*/
tp->t_flags |= TF_DELACK; /*置延迟的ACK标志*/
return;
}
} /*首部预测算法结束*/
/*
* Drop TCP, IP headers and TCP options.
*/
m->m_data += sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr); /*丢弃IP首部,TCP首部*/
m->m_len -= sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
/*
* Calculate amount of space in receive window,
* and then do TCP input processing.
* Receive window is amount of space in rcv queue,
* but not less than advertised window.
*/
{ int win;
win = sbspace(&so->so_rcv); /*接收缓冲区可用空间*/
if (win < 0)
win = 0;
tp->rcv_wnd = max(win, (int)(tp->rcv_adv - tp->rcv_nxt)); /*更新接收窗口大小*/
}
switch (tp->t_state) {
/*
* If the state is LISTEN then ignore segment if it contains an RST.
* If the segment contains an ACK then it is bad and send a RST.
* If it does not contain a SYN then it is not interesting; drop it.
* Don't bother responding if the destination was a broadcast.
* Otherwise initialize tp->rcv_nxt, and tp->irs, select an initial
* tp->iss, and send a segment:
* <SEQ=ISS><ACK=RCV_NXT><CTL=SYN,ACK>
* Also initialize tp->snd_nxt to tp->iss+1 and tp->snd_una to tp->iss.
* Fill in remote peer address fields if not previously specified.
* Enter SYN_RECEIVED state, and process any other fields of this
* segment in this state.
*/
case TCPS_LISTEN: { /*被动打开,期待接收SYN,然后发送SYN、ACK*/
struct mbuf *am;
register struct sockaddr_in *sin;
#ifdef already_done
if (tiflags & TH_RST) /*报文带RST标志,丢弃*/
goto drop;
if (tiflags & TH_ACK) /*报文带ACK标志,丢弃,发送RST*/
goto dropwithreset;
if ((tiflags & TH_SYN) == 0) /*非SYN报文,丢弃*/
goto drop;
#endif
/*
* RFC1122 4.2.3.10, p. 104: discard bcast/mcast SYN
* in_broadcast() should never return true on a received
* packet with M_BCAST not set.
*/
if (m->m_flags & (M_BCAST|M_MCAST) || /*丢弃广播和多播的SYN报文*/
IN_MULTICAST(ntohl(ti->ti_dst.s_addr)))
goto drop;
am = m_get(M_DONTWAIT, MT_SONAME); /*获取一个保存sockaddr_in结构的mbuf*/
if (am == NULL)
goto drop;
am->m_len = sizeof (struct sockaddr_in);
sin = mtod(am, struct sockaddr_in *);
sin->sin_family = AF_INET;
sin->sin_len = sizeof(*sin);
sin->sin_addr = ti->ti_src; /*源IP地址*/
sin->sin_port = ti->ti_sport; /*源端口号*/
bzero((caddr_t)sin->sin_zero, sizeof(sin->sin_zero));
laddr = inp->inp_laddr;
if (inp->inp_laddr.s_addr == INADDR_ANY)
inp->inp_laddr = ti->ti_dst;
if (in_pcbconnect(inp, am)) { /*设置socket的外部地址和外部端口号!*/
inp->inp_laddr = laddr;
(void) m_free(am);
goto drop;
}
(void) m_free(am); /*释放的临时mbuf*/
tp->t_template = tcp_template(tp); /*获取一个TCP IP首部模板*/
if (tp->t_template == 0) {
tp = tcp_drop(tp, ENOBUFS);
dropsocket = 0; /* socket is already gone */
goto drop;
}
if (optp) /*解析LISTEN状态的TCP选项*/*/
tcp_dooptions(tp, optp, optlen, ti,
&ts_present, &ts_val, &ts_ecr);
if (iss)
tp->iss = iss; /*初始的发送序列号*/
else
tp->iss = tcp_iss;
tcp_iss += TCP_ISSINCR/4;
tp->irs = ti->ti_seq; /*初始的接收序列号*/
tcp_sendseqinit(tp); /*初始化发送序列空间*/
/*(tp)->snd_una = (tp)->snd_nxt = (tp)->snd_max = (tp)->snd_up = (tp)->iss*/
tcp_rcvseqinit(tp); /*初始化接收序列空间*/
/*(tp)->rcv_adv = (tp)->rcv_nxt = (tp)->irs + 1*/
tp->t_flags |= TF_ACKNOW; /*立即发送ACK*/
tp->t_state = TCPS_SYN_RECEIVED; /*转移到SYN_RECEIVED状态,发送SYN、ACK*/
tp->t_timer[TCPT_KEEP] = TCPTV_KEEP_INIT; /*启动连接建立定时器*/
dropsocket = 0; /* committed to socket */
tcpstat.tcps_accepts++;
goto trimthenstep6;
}
/*
* If the state is SYN_SENT:
* if seg contains an ACK, but not for our SYN, drop the input.
* if seg contains a RST, then drop the connection.
* if seg does not contain SYN, then drop it.
* Otherwise this is an acceptable SYN segment
* initialize tp->rcv_nxt and tp->irs
* if seg contains ack then advance tp->snd_una
* if SYN has been acked change to ESTABLISHED else SYN_RCVD state
* arrange for segment to be acked (eventually)
* continue processing rest of data/controls, beginning with URG
*/
case TCPS_SYN_SENT: /*主动打开,已经发送过SYN,期待接收SYN、ACK*/
if ((tiflags & TH_ACK) &&
(SEQ_LEQ(ti->ti_ack, tp->iss) ||
SEQ_GT(ti->ti_ack, tp->snd_max))) /*无效的ACK,发送RST*/
goto dropwithreset;
if (tiflags & TH_RST) { /*带RST标志*/
if (tiflags & TH_ACK) /*又带ACK标志,对端拒绝连接,丢弃连接*/
tp = tcp_drop(tp, ECONNREFUSED);
goto drop;
}
if ((tiflags & TH_SYN) == 0) /*不带SYN标志,丢弃*/
goto drop;
if (tiflags & TH_ACK) { /*带ACK标志*/
tp->snd_una = ti->ti_ack; /*更新snd_una指针*/
if (SEQ_LT(tp->snd_nxt, tp->snd_una)) /*更新snd_nxt指针*/
tp->snd_nxt = tp->snd_una;
}
tp->t_timer[TCPT_REXMT] = 0;
tp->irs = ti->ti_seq; /*初始的接收序列号*/*/
tcp_rcvseqinit(tp); /*初始化接收序列空间*/
/*(tp)->rcv_adv = (tp)->rcv_nxt = (tp)->irs + 1*/
tp->t_flags |= TF_ACKNOW; /*立即发送ACK*/*/
if (tiflags & TH_ACK && SEQ_GT(tp->snd_una, tp->iss)) { /*收到的ACK已经确认了发送的SYN*/
tcpstat.tcps_connects++;
soisconnected(so); /*TCP连接已建立*/
tp->t_state = TCPS_ESTABLISHED; /*状态转移到ESTABLISHED!*/
/* Do window scaling on this connection? */
if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) ==
(TF_RCVD_SCALE|TF_REQ_SCALE)) { /*记录窗口缩放因子*/
tp->snd_scale = tp->requested_s_scale;
tp->rcv_scale = tp->request_r_scale;
}
(void) tcp_reass(tp, (struct tcpiphdr *)0, /*由于数据可能在连接未建立之前到达,调用tcp_reass将数据放入接收缓冲区*/
(struct mbuf *)0);
/*
* if we didn't have to retransmit the SYN,
* use its rtt as our initial srtt & rtt var.
*/
if (tp->t_rtt) /*如果第一个SYN正被计时,计算第一个RTO*/
tcp_xmit_timer(tp, tp->t_rtt);
} else
tp->t_state = TCPS_SYN_RECEIVED; /*同时打开*/
trimthenstep6:
/*
* Advance ti->ti_seq to correspond to first data byte.
* If data, trim to stay within window,
* dropping FIN if necessary.
*/
ti->ti_seq++; /*SYN标志占用一个序号,SYN报文可能带有数据?*/
if (ti->ti_len > tp->rcv_wnd) { /*丢弃超过接收窗口的数据*/
todrop = ti->ti_len - tp->rcv_wnd;
m_adj(m, -todrop);
ti->ti_len = tp->rcv_wnd;
tiflags &= ~TH_FIN;
tcpstat.tcps_rcvpackafterwin++;
tcpstat.tcps_rcvbyteafterwin += todrop;
}
tp->snd_wl1 = ti->ti_seq - 1; /*会强制更新3个窗口变量*/
tp->rcv_up = ti->ti_seq;
goto step6; /*调到879行*/
}
/*
* States other than LISTEN or SYN_SENT.
* First check timestamp, if present.
* Then check that at least some bytes of segment are within
* receive window. If segment begins before rcv_nxt,
* drop leading data (and SYN); if nothing left, just ack.
*
* RFC 1323 PAWS: If we have a timestamp reply on this segment
* and it's less than ts_recent, drop it.
*/
if (ts_present && (tiflags & TH_RST) == 0 && tp->ts_recent &&
TSTMP_LT(ts_val, tp->ts_recent)) { /*重复的报文段*/
/* Check to see if ts_recent is over 24 days old. */
if ((int)(tcp_now - tp->ts_recent_age) > TCP_PAWS_IDLE) { /*时间戳已经回绕*/
/*
* Invalidate ts_recent. If this segment updates
* ts_recent, the age will be reset later and ts_recent
* will get a valid value. If it does not, setting
* ts_recent to zero will at least satisfy the
* requirement that zero be placed in the timestamp
* echo reply when ts_recent isn't valid. The
* age isn't reset until we get a valid ts_recent
* because we don't want out-of-order segments to be
* dropped when ts_recent is old.
*/
tp->ts_recent = 0; /*使记录的对端的时间戳无效*/
} else { /*更新统计信息*/
tcpstat.tcps_rcvduppack++;
tcpstat.tcps_rcvdupbyte += ti->ti_len;
tcpstat.tcps_pawsdrop++;
goto dropafterack;
}
}
todrop = tp->rcv_nxt - ti->ti_seq; /*超过接收窗口左边沿的字节数*/
if (todrop > 0) {
if (tiflags & TH_SYN) { /*重复的SYN?*/
tiflags &= ~TH_SYN;
ti->ti_seq++;
if (ti->ti_urp > 1)
ti->ti_urp--;
else
tiflags &= ~TH_URG;
todrop--;
}
if (todrop > ti->ti_len ||
todrop == ti->ti_len && (tiflags & TH_FIN) == 0 ) { //完全重复的报文段
tiflags &= ~TH_FIN;
tp->t_flags |= TF_ACKNOW; //立即发送ACK
todrop = ti->ti_len;
tcpstat.tcps_rcvduppack++;
tcpstat.tcps_rcvdupbyte += ti->ti_len;
} else { //部分重复
tcpstat.tcps_rcvpartduppack++;
tcpstat.tcps_rcvpartdupbyte += todrop;
}
m_adj(m, todrop); /*丢弃重复的数据,修正TCP首部中相关字段*/
ti->ti_seq += todrop;
ti->ti_len -= todrop;
if (ti->ti_urp > todrop)
ti->ti_urp -= todrop;
else {
tiflags &= ~TH_URG;
ti->ti_urp = 0;
}
}
/*
* If new data are received on a connection after the
* user processes are gone, then RST the other end.
*/
if ((so->so_state & SS_NOFDREF) &&
tp->t_state > TCPS_CLOSE_WAIT && ti->ti_len) { /*应用程序关闭了连接,如果收到数据,则发送RST*/
tp = tcp_close(tp);
tcpstat.tcps_rcvafterclose++;
goto dropwithreset;
}
/*
* If segment ends after window, drop trailing data
* (and PUSH and FIN); if nothing left, just ACK.
*/
todrop = (ti->ti_seq+ti->ti_len) - (tp->rcv_nxt+tp->rcv_wnd); /*超过接收窗口右边沿的字节数*/
if (todrop > 0) {
tcpstat.tcps_rcvpackafterwin++;
if (todrop >= ti->ti_len) { /*完全重复的报文段*/
tcpstat.tcps_rcvbyteafterwin += ti->ti_len;
/*
* If a new connection request is received
* while in TIME_WAIT, drop the old connection
* and start over if the sequence numbers
* are above the previous ones.
*/
if (tiflags & TH_SYN &&
tp->t_state == TCPS_TIME_WAIT &&
SEQ_GT(ti->ti_seq, tp->rcv_nxt)) { /*重建连接??*/
iss = tp->snd_nxt + TCP_ISSINCR;
tp = tcp_close(tp);
goto findpcb;
}
/*
* If window is closed can only take segments at
* window edge, and have to drop data and PUSH from
* incoming segments. Continue processing, but
* remember to ack. Otherwise, drop segment
* and ack.
*/
if (tp->rcv_wnd == 0 && ti->ti_seq == tp->rcv_nxt) { /*收到了窗口探测报文,立即发送ACK*/
tp->t_flags |= TF_ACKNOW;
tcpstat.tcps_rcvwinprobe++;
} else
goto dropafterack;
} else
tcpstat.tcps_rcvbyteafterwin += todrop;
m_adj(m, -todrop); /*丢弃重复的数据,修正TCP首部中相关字段*/
ti->ti_len -= todrop;
tiflags &= ~(TH_PUSH|TH_FIN);
}
if (ts_present && TSTMP_GEQ(ts_val, tp->ts_recent) &&
SEQ_LEQ(ti->ti_seq, tp->last_ack_sent)) { //哪个时间戳需要回显,正确的算法!!!!
tp->ts_recent_age = tcp_now; //最近一次ts_recent被更新的时间戳
tp->ts_recent = ts_val; //对端发送的最新的有效时间戳
}
/*
* If the RST bit is set examine the state:
* SYN_RECEIVED STATE:
* If passive open, return to LISTEN state.
* If active open, inform user that connection was refused.
* ESTABLISHED, FIN_WAIT_1, FIN_WAIT2, CLOSE_WAIT STATES:
* Inform user that connection was reset, and close tcb.
* CLOSING, LAST_ACK, TIME_WAIT STATES
* Close the tcb.
*/
if (tiflags&TH_RST) switch (tp->t_state) { /*处理RST标志*/
case TCPS_SYN_RECEIVED:
so->so_error = ECONNREFUSED;
goto close;
case TCPS_ESTABLISHED:
case TCPS_FIN_WAIT_1:
case TCPS_FIN_WAIT_2:
case TCPS_CLOSE_WAIT:
so->so_error = ECONNRESET;
close:
tp->t_state = TCPS_CLOSED;
tcpstat.tcps_drops++;
tp = tcp_close(tp);
goto drop;
case TCPS_CLOSING:
case TCPS_LAST_ACK:
case TCPS_TIME_WAIT:
tp = tcp_close(tp);
goto drop;
}
/*
* If a SYN is in the window, then this is an
* error and we send an RST and drop the connection.
*/
if (tiflags & TH_SYN) {
tp = tcp_drop(tp, ECONNRESET);
goto dropwithreset;
}
/*
* If the ACK bit is off we drop the segment and return.
*/
if ((tiflags & TH_ACK) == 0)
goto drop;
/*
* Ack processing.
*/
switch (tp->t_state) { /*处理ACK标志*/
/*
* In SYN_RECEIVED state if the ack ACKs our SYN then enter
* ESTABLISHED state and continue processing, otherwise
* send an RST.
*/
case TCPS_SYN_RECEIVED: /*服务端,期待收到SYN、ACK,然后发送ACK*/
if (SEQ_GT(tp->snd_una, ti->ti_ack) ||
SEQ_GT(ti->ti_ack, tp->snd_max)) /*无效的ACK,发送RST*/
goto dropwithreset;
tcpstat.tcps_connects++;
soisconnected(so); /*连接已经建立*/
tp->t_state = TCPS_ESTABLISHED; /*状态转移到ESTABLISHED!*/
/* Do window scaling? */
if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) ==
(TF_RCVD_SCALE|TF_REQ_SCALE)) { /*记录窗口缩放因子*/
tp->snd_scale = tp->requested_s_scale;
tp->rcv_scale = tp->request_r_scale;
}
(void) tcp_reass(tp, (struct tcpiphdr *)0, (struct mbuf *)0); /*将数据加入到接收缓冲区*/
tp->snd_wl1 = ti->ti_seq - 1; /*会强制更新3个窗口变量*/
/* fall into ... */
/*
* In ESTABLISHED state: drop duplicate ACKs; ACK out of range
* ACKs. If the ack is in the range
* tp->snd_una < ti->ti_ack <= tp->snd_max
* then advance tp->snd_una to ti->ti_ack and drop
* data from the retransmission queue. If this ACK reflects
* more up to date window information we update our window information.
*/
case TCPS_ESTABLISHED:
case TCPS_FIN_WAIT_1:
case TCPS_FIN_WAIT_2:
case TCPS_CLOSE_WAIT:
case TCPS_CLOSING:
case TCPS_LAST_ACK:
case TCPS_TIME_WAIT:
/*快速重传和快速恢复算法 !!!*/
if (SEQ_LEQ(ti->ti_ack, tp->snd_una)) { /*接收到重复的ACK,即1. ACK未确认数据*/
if (ti->ti_len == 0 && tiwin == tp->snd_wnd) { /*2. 报文没有携带数据 3. 通告的窗口大小未变*/
tcpstat.tcps_rcvdupack++;
/*
* If we have outstanding data (other than
* a window probe), this is a completely
* duplicate ack (ie, window info didn't
* change), the ack is the biggest we've
* seen and we've seen exactly our rexmt
* threshhold of them, assume a packet
* has been dropped and retransmit it.
* Kludge snd_nxt & the congestion
* window so we send only this one
* packet.
*
* We know we're losing at the current
* window size so do congestion avoidance
* (set ssthresh to half the current window
* and pull our congestion window back to
* the new ssthresh).
*
* Dup acks mean that packets have left the
* network (they're now cached at the receiver)
* so bump cwnd by the amount in the receiver
* to keep a constant cwnd packets in the
* network.
*/
if (tp->t_timer[TCPT_REXMT] == 0 ||
ti->ti_ack != tp->snd_una)
tp->t_dupacks = 0; /*重复的ACK计算清零*/
else if (++tp->t_dupacks == tcprexmtthresh) { /*4. 重传定时器已启动 5.
确认字段是TCP收到的最大确认序列号 可确定是完全重复的ACK。
如果重复的ACK次数为3*/
tcp_seq onxt = tp->snd_nxt;
/*令慢启动门限为当前拥塞窗口的一半,最小值为两个报文段*/
u_int win =
min(tp->snd_wnd, tp->snd_cwnd) / 2 /
tp->t_maxseg;
if (win < 2)
win = 2;
tp->snd_ssthresh = win * tp->t_maxseg;
tp->t_timer[TCPT_REXMT] = 0; /*关闭重传定时器*/
tp->t_rtt = 0; /*关闭报文段计时*/
tp->snd_nxt = ti->ti_ack;
tp->snd_cwnd = tp->t_maxseg; /*拥塞窗口设置为一个报文段大小*/
(void) tcp_output(tp); /*重传报文段*/
tp->snd_cwnd = tp->snd_ssthresh +
tp->t_maxseg * tp->t_dupacks; /*重传后,拥塞窗口大小为慢启动门限加上3倍的报文段大小*/
if (SEQ_GT(onxt, tp->snd_nxt))
tp->snd_nxt = onxt;
goto drop;
} else if (tp->t_dupacks > tcprexmtthresh) { /*重复的ACK超过3*/
tp->snd_cwnd += tp->t_maxseg; /*增大拥塞窗口大小增大一个报文段大小*/
(void) tcp_output(tp); /*发送数据*/
goto drop;
}
} else /*重复的ACK计算清零*/
tp->t_dupacks = 0;
break;
}
/*
* If the congestion window was inflated to account
* for the other side's cached packets, retract it.
*/
if (tp->t_dupacks > tcprexmtthresh && /*非重复ACK,快速重传结束*/
tp->snd_cwnd > tp->snd_ssthresh)
tp->snd_cwnd = tp->snd_ssthresh;
tp->t_dupacks = 0; /*重复的ACK计算清零*/
acked = ti->ti_ack - tp->snd_una; /*ACK确认的字节数*/
tcpstat.tcps_rcvackpack++;
tcpstat.tcps_rcvackbyte += acked;
/*
* If we have a timestamp reply, update smoothed
* round trip time. If no timestamp is present but
* transmit timer is running and timed sequence
* number was acked, update smoothed round trip time.
* Since we now have an rtt measurement, cancel the
* timer backoff (cf., Phil Karn's retransmit alg.).
* Recompute the initial retransmit timer.
*/
/*计算RTT,根据RTT计算RTO*/
if (ts_present)
tcp_xmit_timer(tp, tcp_now-ts_ecr+1);
else if (tp->t_rtt && SEQ_GT(ti->ti_ack, tp->t_rtseq))
tcp_xmit_timer(tp,tp->t_rtt);
/*
* If all outstanding data is acked, stop retransmit
* timer and remember to restart (more output or persist).
* If there is more data to be acked, restart retransmit
* timer, using current (possibly backed-off) value.
*/
if (ti->ti_ack == tp->snd_max) { /*ACK已经确认了所有的数据,关闭重传定时器*/
tp->t_timer[TCPT_REXMT] = 0;
needoutput = 1; /*发送窗口已增大,发送更多数据!*/
} else if (tp->t_timer[TCPT_PERSIST] == 0) /*非持续状态,启动重传定时器*/
tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
/*
* When new data is acked, open the congestion window.
* If the window gives us less than ssthresh packets
* in flight, open exponentially (maxseg per packet).
* Otherwise open linearly: maxseg per window
* (maxseg * (maxseg / cwnd) per packet).
*/
{ /*慢启动和拥塞避免,更新拥塞窗口!!*/
register u_int cw = tp->snd_cwnd;
register u_int incr = tp->t_maxseg;
if (cw > tp->snd_ssthresh) /*正在进行拥塞避免,否则正在进行慢启动*/
incr = incr * incr / cw;
tp->snd_cwnd = min(cw + incr, TCP_MAXWIN<<tp->snd_scale); /*更新拥塞窗口大小*/
}
if (acked > so->so_snd.sb_cc) { /*发送缓冲区所有的数据被确认,同时说明发送的FIN已被确认*/
tp->snd_wnd -= so->so_snd.sb_cc; /*更新发送窗口大小*/
sbdrop(&so->so_snd, (int)so->so_snd.sb_cc); /*丢弃发送缓冲区中的数据*/
ourfinisacked = 1;
} else {
sbdrop(&so->so_snd, acked); /*丢弃已被确认的数据*/
tp->snd_wnd -= acked; /*更新发送窗口大小*/
ourfinisacked = 0;
}
if (so->so_snd.sb_flags & SB_NOTIFY) /*唤醒在发送缓冲区等待的进程*/*/
sowwakeup(so);
tp->snd_una = ti->ti_ack; /*更新snd_una指针*/
if (SEQ_LT(tp->snd_nxt, tp->snd_una)) /*更新snd_nxt指针*/
tp->snd_nxt = tp->snd_una;
switch (tp->t_state) { /*特殊状态下的ACK处理*/
/*
* In FIN_WAIT_1 STATE in addition to the processing
* for the ESTABLISHED state if our FIN is now acknowledged
* then enter FIN_WAIT_2.
*/
case TCPS_FIN_WAIT_1: /*主动关闭,已发送过FIN,期待接收ACK*/
if (ourfinisacked) { /*发送的FIN被确认*/
/*
* If we can't receive any more
* data, then closing user can proceed.
* Starting the timer is contrary to the
* specification, but if we don't get a FIN
* we'll hang forever.
*/
if (so->so_state & SS_CANTRCVMORE) { /*当应用程序完全关闭了连接时,启动FIN_WAIT_2定时器*/
soisdisconnected(so);
tp->t_timer[TCPT_2MSL] = tcp_maxidle;
}
tp->t_state = TCPS_FIN_WAIT_2; /*转移到FIN_WAIT_2状态*/
}
break;
/*
* In CLOSING STATE in addition to the processing for
* the ESTABLISHED state if the ACK acknowledges our FIN
* then enter the TIME-WAIT state, otherwise ignore
* the segment.
*/
case TCPS_CLOSING: /*同时关闭??*/
if (ourfinisacked) {
tp->t_state = TCPS_TIME_WAIT;
tcp_canceltimers(tp);
tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
soisdisconnected(so);
}
break;
/*
* In LAST_ACK, we may still be waiting for data to drain
* and/or to be acked, as well as for the ack of our FIN.
* If our FIN is now acknowledged, delete the TCB,
* enter the closed state and return.
*/
case TCPS_LAST_ACK: /*服务端,已发送过FIN,等待接收最后一个ACK*/
if (ourfinisacked) { /*FIN已被确认,关闭连接*/
tp = tcp_close(tp);
goto drop;
}
break;
/*
* In TIME_WAIT state the only thing that should arrive
* is a retransmission of the remote FIN. Acknowledge
* it and restart the finack timer.
*/
case TCPS_TIME_WAIT: /*如果对端发送的FIN丢失,对端会重传FIN(带有ACK),那么丢弃报文并重传ACK*/
tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL; /*重新启动TIME_WAIT定时器*/
goto dropafterack;
}
}
step6:
/*
* Update window information.
* Don't look at window if no ACK: TAC's send garbage on first SYN.
*/
/*snd_wl1记录最后接收报文段的序号(对端的序号),用于更新发送窗口*/
/*snd_wl2记录最后接收报文段的确认序号(本端的序号),用于更新发送窗口*/
if ((tiflags & TH_ACK) && /*ACK标志置位*/
(SEQ_LT(tp->snd_wl1, ti->ti_seq) || tp->snd_wl1 == ti->ti_seq && /*报文携带了新数据*/
(SEQ_LT(tp->snd_wl2, ti->ti_ack) || /*未携带新数据,但确认了数据*/
tp->snd_wl2 == ti->ti_ack && tiwin > tp->snd_wnd))) { /*未携带新数据,未确认数据,但通告窗口大于当前发送窗口*/
/* keep track of pure window updates */
if (ti->ti_len == 0 &&
tp->snd_wl2 == ti->ti_ack && tiwin > tp->snd_wnd) /*窗口更新报文统计*/
tcpstat.tcps_rcvwinupd++;
tp->snd_wnd = tiwin; /*更新发送窗口(由对端通告的)大小*/
tp->snd_wl1 = ti->ti_seq; /*更新snd_wl1指针*/
tp->snd_wl2 = ti->ti_ack; /*更新snd_wl2指针*/
if (tp->snd_wnd > tp->max_sndwnd)
tp->max_sndwnd = tp->snd_wnd;
needoutput = 1; /*发送窗口已增大,发送更多数据!*/
}
/*
* If no out of band data is expected,
* pull receive urgent pointer along
* with the receive window.
*/
if (SEQ_GT(tp->rcv_nxt, tp->rcv_up))
tp->rcv_up = tp->rcv_nxt;
dodata: /* XXX */
/*
* Process the segment text, merging it into the TCP sequencing queue,
* and arranging for acknowledgment of receipt if necessary.
* This process logically involves adjusting tp->rcv_wnd as data
* is presented to the user (this happens in tcp_usrreq.c,
* case PRU_RCVD). If a FIN has already been received on this
* connection then we just ignore the text.
*/
if ((ti->ti_len || (tiflags&TH_FIN)) &&
TCPS_HAVERCVDFIN(tp->t_state) == 0) {
TCP_REASS(tp, ti, m, so, tiflags); /*调用宏TCP_REASS处理数据*/
/*
* Note the amount of data that peer has sent into
* our window, in order to estimate the sender's
* buffer size.
*/
len = so->so_rcv.sb_hiwat - (tp->rcv_adv - tp->rcv_nxt);
} else {
m_freem(m);
tiflags &= ~TH_FIN;
}
/*
* If FIN is received ACK the FIN and let the user know
* that the connection is closing.
*/
if (tiflags & TH_FIN) { /*处理FIN标志*/
if (TCPS_HAVERCVDFIN(tp->t_state) == 0) {
socantrcvmore(so); /*不能再接收数据*/
tp->t_flags |= TF_ACKNOW; /*立即发送ACK*/
tp->rcv_nxt++;
}
switch (tp->t_state) {
/*
* In SYN_RECEIVED and ESTABLISHED STATES
* enter the CLOSE_WAIT state.
*/
case TCPS_SYN_RECEIVED:
case TCPS_ESTABLISHED:
tp->t_state = TCPS_CLOSE_WAIT;
break;
/*
* If still in FIN_WAIT_1 STATE FIN has not been acked so
* enter the CLOSING state.
*/
case TCPS_FIN_WAIT_1:
tp->t_state = TCPS_CLOSING;
break;
/*
* In FIN_WAIT_2 state enter the TIME_WAIT state,
* starting the time-wait timer, turning off the other
* standard timers.
*/
case TCPS_FIN_WAIT_2: /*进入TIME_WAIT状态*/
tp->t_state = TCPS_TIME_WAIT;
tcp_canceltimers(tp);
tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
soisdisconnected(so); /*连接已断开*/
break;
/*
* In TIME_WAIT state restart the 2 MSL time_wait timer.
*/
case TCPS_TIME_WAIT:
tp->t_timer[TCPT_2MSL] = 2 * TCPTV_MSL;
break;
}
}
/*
* Return any desired output.
*/
if (needoutput || (tp->t_flags & TF_ACKNOW)) /*如果要立即发送ACK或有更多的数据要发送,调用tcp_output*/
(void) tcp_output(tp);
return;
dropafterack: /*发送ACK,然后丢弃报文*/
/*
* Generate an ACK dropping incoming segment if it occupies
* sequence space, where the ACK reflects our state.
*/
if (tiflags & TH_RST)
goto drop;
m_freem(m);
tp->t_flags |= TF_ACKNOW;
(void) tcp_output(tp);
return;
dropwithreset: /*发送RST,然后丢弃报文*/
/*
* Generate a RST, dropping incoming segment.
* Make ACK acceptable to originator of segment.
* Don't bother to respond if destination was broadcast/multicast.
*/
if ((tiflags & TH_RST) || m->m_flags & (M_BCAST|M_MCAST) ||
IN_MULTICAST(ntohl(ti->ti_dst.s_addr)))
goto drop;
if (tiflags & TH_ACK) /*带有ACK,RST报文序号字段是接收到的确认序列号,确认序列号字段是0*/
tcp_respond(tp, ti, m, (tcp_seq)0, ti->ti_ack, TH_RST);
else {
if (tiflags & TH_SYN) /*如果到达的SYN请求与不存在的服务器建立连接,会执行这段代码*/
ti->ti_len++;
tcp_respond(tp, ti, m, ti->ti_seq+ti->ti_len, (tcp_seq)0, /*不带有ACK,RST报文带有ACK标志,
序号字段是0,确认序列号字段是接收到的序号字段*/
TH_RST|TH_ACK);
}
/* destroy temporarily created socket */
if (dropsocket)
(void) soabort(so);
return;
drop: /*直接丢弃报文*/
/*
* Drop space held by incoming segment and return.
*/
m_freem(m);
/* destroy temporarily created socket */
if (dropsocket)
(void) soabort(so);
return;
}
sonewconn函数为新连接创建一个socket,它的代码如下:
/*
* When an attempt at a new connection is noted on a socket
* which accepts connections, sonewconn is called. If the
* connection is possible (subject to space constraints, etc.)
* then we allocate a new structure, propoerly linked into the
* data structure of the original socket, and return this.
* Connstatus may be 0, or SO_ISCONFIRMING, or SO_ISCONNECTED.
*
* Currently, sonewconn() is defined as sonewconn1() in socketvar.h
* to catch calls that are missing the (new) second parameter.
*/
struct socket *
sonewconn(head, connstatus)
register struct socket *head;
int connstatus;
{
register struct socket *so;
int soqueue = connstatus ? 1 : 0;
if (head->so_qlen + head->so_q0len > 3 * head->so_qlimit / 2) /*不能再接收一个新连接,返回NULL*/
return ((struct socket *)0);
MALLOC(so, struct socket *, sizeof(*so), M_SOCKET, M_DONTWAIT); /*分配一个socket结构*/
if (so == NULL)
return ((struct socket *)0);
bzero((caddr_t)so, sizeof(*so));
so->so_type = head->so_type;
so->so_options = head->so_options &~ SO_ACCEPTCONN; /*去掉SO_ACCEPTCONN标志位,继承监听socket的选项*/
so->so_linger = head->so_linger;
so->so_state = head->so_state | SS_NOFDREF; /*该socket没有与描述符关联*/
so->so_proto = head->so_proto;
so->so_timeo = head->so_timeo;
so->so_pgid = head->so_pgid;
(void) soreserve(so, head->so_snd.sb_hiwat, head->so_rcv.sb_hiwat); /*设置socket的接收缓冲区和发送缓冲区大小*/
soqinsque(head, so, soqueue); /*将socket加入等待完成三次握手的连接队列*/
if ((*so->so_proto->pr_usrreq)(so, PRU_ATTACH, /*以PRU_ATTACH命令调用tcp_usrreq函数*/
(struct mbuf *)0, (struct mbuf *)0, (struct mbuf *)0)) {
(void) soqremque(so, soqueue);
(void) free((caddr_t)so, M_SOCKET);
return ((struct socket *)0);
}
if (connstatus) { /*连接已经建立*/
sorwakeup(head);
wakeup((caddr_t)&head->so_timeo); /*唤醒等待在accept系统调用的进程!*/
so->so_state |= connstatus;
}
return (so);
}
处理TCP连接请求的过程如下:
处理TCP选项的函数tcp_dooptions的代码如下:
void
tcp_dooptions(tp, cp, cnt, ti, ts_present, ts_val, ts_ecr)
struct tcpcb *tp;
u_char *cp;
int cnt;
struct tcpiphdr *ti;
int *ts_present;
u_long *ts_val, *ts_ecr;
{
u_short mss;
int opt, optlen;
for (; cnt > 0; cnt -= optlen, cp += optlen) {
opt = cp[0];
if (opt == TCPOPT_EOL) /*选项结束*/
break;
if (opt == TCPOPT_NOP)
optlen = 1;
else {
optlen = cp[1];
if (optlen <= 0)
break;
}
switch (opt) {
default:
continue;
case TCPOPT_MAXSEG: /*MSS选项*/
if (optlen != TCPOLEN_MAXSEG)
continue;
if (!(ti->ti_flags & TH_SYN))
continue;
bcopy((char *) cp + 2, (char *) &mss, sizeof(mss));
NTOHS(mss);
(void) tcp_mss(tp, mss); /* 调用tcp_mss函数处理 */
break;
case TCPOPT_WINDOW: /*窗口缩放因子选项*/
if (optlen != TCPOLEN_WINDOW)
continue;
if (!(ti->ti_flags & TH_SYN))
continue;
tp->t_flags |= TF_RCVD_SCALE;
tp->requested_s_scale = min(cp[2], TCP_MAX_WINSHIFT);
break;
case TCPOPT_TIMESTAMP: /*时间戳选项*/
if (optlen != TCPOLEN_TIMESTAMP)
continue;
*ts_present = 1;
bcopy((char *)cp + 2, (char *) ts_val, sizeof(*ts_val));
NTOHL(*ts_val);
bcopy((char *)cp + 6, (char *) ts_ecr, sizeof(*ts_ecr));
NTOHL(*ts_ecr);
/*
* A timestamp received in a SYN makes
* it ok to send timestamp requests and replies.
*/
if (ti->ti_flags & TH_SYN) { /*SYN报文的时间戳*/
tp->t_flags |= TF_RCVD_TSTMP;
tp->ts_recent = *ts_val;
tp->ts_recent_age = tcp_now;
}
break;
}
}
}
处理接收数据的TCP_REASS宏定义如下:
/*
* Insert segment ti into reassembly queue of tcp with
* control block tp. Return TH_FIN if reassembly now includes
* a segment with FIN. The macro form does the common case inline
* (segment is the next to be received on an established connection,
* and the queue is empty), avoiding linkage into and removal
* from the queue and repetition of various conversions.
* Set DELACK for segments received in order, but ack immediately
* when segments are out of order (so fast retransmit can work). 2103
*/
#define TCP_REASS(tp, ti, m, so, flags) { \
if ((ti)->ti_seq == (tp)->rcv_nxt && \
(tp)->seg_next == (struct tcpiphdr *)(tp) && \
(tp)->t_state == TCPS_ESTABLISHED) { \
tp->t_flags |= TF_DELACK; \
(tp)->rcv_nxt += (ti)->ti_len; \
flags = (ti)->ti_flags & TH_FIN; \
tcpstat.tcps_rcvpack++;\
tcpstat.tcps_rcvbyte += (ti)->ti_len;\
sbappend(&(so)->so_rcv, (m)); \
sorwakeup(so); \
} else { \
(flags) = tcp_reass((tp), (ti), (m)); \
tp->t_flags |= TF_ACKNOW; \
} \
}
如果到达的数据顺序正确 ,置上延迟ACK标志,更新rcv_nxt指针,并把数据添加到socket的接收缓冲区中。如果数据顺序错误,宏会调用tcp_reass函数,把数据加入到连接的重组队列中(新到数据有可能填充队列中的缺口,从而将已排队的数据添加到socket的接收缓冲区中)。